I am thrilled to announce that the National Science Foundation has funded our Science and Technology Center, Learning the Earth with Artificial intelligence and Physics (LEAP). NSF’s announcement is here.
This Center will use existing machine learning to improve climate model fidelity, and to develop better data products for climate model validation. We will also develop new techniques in machine learning for application to the smaller and sparser datasets that are common in geoscience. More details on the science can be found in this Q&A with me and Professor Carl Vondrick.
The Center will be led by Professor Pierre Gentine of Columbia Earth and Environmental Engineering and Earth and Environmental Sciences. I will be Deputy Director. Many colleagues at Lamont-Doherty Earth Observatory, in Engineering, and in Arts and Sciences are on our team. We have partners at several other institutions across the country. A key one is NCAR since it is their Community Earth System Model (CESM) that is our primary model for implementation of improved algorithms. Columbia’s website announcement offers more information and introduces the full team.
I look forward to this exciting project, and to sharing with everyone the LEAPs in climate modeling and in machine learning that we aim to produce.
– Galen McKinley